[R] Inverting a scale(X)
G'day, All. I have been trying to trackdown a problem in my R analysis script. I perform a scale() operation on a matrix then do further work. Is there any way of inverting the scale() such that sX - scale(X) Xprime - inv.scale(x); # does inv.scale exist? resulting in Xprime_{ij} == X_{ij} where Xprime_{ij} \in R There must be some way of doing it but I'm such a newb that I haven't been able to find it. Thanks Godfrey __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Inverting a scale(X)
On 2010-07-03 0:05, Godfrey van der Linden wrote: G'day, All. I have been trying to trackdown a problem in my R analysis script. I perform a scale() operation on a matrix then do further work. Is there any way of inverting the scale() such that sX- scale(X) Xprime- inv.scale(x); # does inv.scale exist? resulting in Xprime_{ij} == X_{ij} where Xprime_{ij} \in R There must be some way of doing it but I'm such a newb that I haven't been able to find it. Thanks Godfrey If your sX hasn't lost the scaled:center and scaled:scale attributes that it got from the scale() operation, then you can just reverse the scaling procedure using those. Multiply columns by the scale attribute, then add the center attribute. Something like: MN - attr(sx, scaled:center) SD - attr(sx, scaled:scale) Xprime - t(apply(sx, 1, function(x){x * SD + MN})) If the attributes have been lost by your further work, then I'm afraid you're out of luck. -Peter Ehlers __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Assigning entries to categories
Thanks for your help! You are right it is not one-to-one assigned that would be indeed very easy... its more like assigning 1000 entries to 60 categories... Unfortunately, the ?match and ?merge did not help me a lot... I am a newbie to such programming stuff in R. It would be great if you could help me again to set this up. -- View this message in context: http://r.789695.n4.nabble.com/Assigning-entries-to-categories-tp2272697p2277140.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Best way to compute a sum
Although it does not apply to your series and is impractical, it seems to me that the most accurate algorithm might be to add all the rational numbers whose sum and components can be represented without error in binary first, ie 2.5 + .5 or 1/16 + 1/16 + 1/8. You could also get very clever and investigate a sum that should have an exact binary representation when the individual components do not, ie .1 + .2 + .2 = .5 and correct the sum. Roger -- View this message in context: http://r.789695.n4.nabble.com/Best-way-to-compute-a-sum-tp2267566p2277096.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] conditional dataframe search and find
After some processing... ct.df- data.frame(time,conc) ct.df gives time conc 1 0 164.495456 2 1 133.671185 3 2 108.622975 4 3 88.268468 5 4 71.728126 6 5 58.287225 7 6 47.364971 8 7 38.489403 9 8 31.276998 109 25.416103 11 10 20.653462 12 11 16.783276 13 12 13.638313 14 13 11.082674 15 14 9.005927 16 15 7.318336 - Ignored: 17 16 5.946977 18 17 4.832592 19 18 3.927028 20 19 3.191155 21 20 2.593175 22 21 2.107248 23 22 1.712378 24 23 1.391501 25 24 1.130752 I need to find the time when conc 25. I can read it off the table but I am looking for a programmatic solution. Thanks. Oscar -- Oscar A. Linares, MD Clinical Assistant Professor of Medicine Department of Medicine University of Toledo College of Medicine Toledo, Ohio 43606-3390 Attending Physician The Detroit Medical Center (DMC) Harper University Hospital Wayne State University School of Medicine Detroit, Michigan 48201 Director Translational Pharmacokinetics Pharmacogenomics Unit, La Plaisance Bay, Bolles Harbor, MI Medical Director Monroe Pain Center Monroe, MI 48162 (http:www.monroepaincenter.com) Phone (734) 240-8400 Cell (734) 637-7997 Fax (734) 243-6254 oalinare...@gmail.com [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Error in solve.default
Hello I use c++ program, what call R-project to solve matrix multiplications. Some times, I get an error in R: Error in solve.default(V, R) : system is computationally singular: reciprocal condition number = 2.20828e-19 Execution halted After that, the program crash. The code, what i execute, is: R.assign(arrMeans, string(R)); R.assign(arrCov, string(V)); SEXP ans; int iRet = R.parseEval(solve(V, R), ans); Where R - vector of n size and V - matrix of n,n. Can anyone tell me, what this error means? I have check my matrix and didn't found this number. Is it because matrix too big? How can I, at last, avoid program crashing in R (it crash inside parseEval function)? WBR, Dima [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] help on bar chart
Hey guys, This is the bar chart that I am working on: library(lattice); data - data.frame( X1 = c(2300, 1300, 1300, 450), X2 = c(2110, 2220, 1100, 660), Y = factor(c(sample1, sample2, sample3, sample4)) ); barchart( Y ~ X1 + X2, data, stack = TRUE, horiz = TRUE, lwd = 1.5, xlab = expression(bold(Sample size)), col = colors()[c(24,1)], xlim = c(0,5000), xat = seq(0,5000,1000) ); I wanted to make a bar chart that has hatching lines inside the bar: with sample 2 and 4 having vertical lines and sample 1 and 3 having horizontal lines, like the following (I kind of photoshopped the image to demonstrate what I wanted it to look like): http://r.789695.n4.nabble.com/file/n2277107/test.png Anyone knows how I can add hatching to the bar charts? Thanks very much for your time!!! -- View this message in context: http://r.789695.n4.nabble.com/help-on-bar-chart-tp2277107p2277107.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] conditional dataframe search and find
ct.df[ct.df$conc 25,time] [1] 10 11 12 13 14 15 ct.df[ct.df$conc 25,time][1] df[df$conc 25,time][1] [1] 10 See also help(order) if conc is not ordered. On 02/07/10 22:50, oscar linares wrote: time conc 1 0 164.495456 2 1 133.671185 3 2 108.622975 4 3 88.268468 5 4 71.728126 6 5 58.287225 7 6 47.364971 8 7 38.489403 9 8 31.276998 109 25.416103 11 10 20.653462 12 11 16.783276 13 12 13.638313 14 13 11.082674 15 14 9.005927 16 15 7.318336 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Double Integration
There used to be an adapt package with an integrate function (I inverted the function/package name by mistake) in CRAN but it has been removed. Anyone knows why? Christos CC: argch...@hotmail.com; sarah_sanche...@yahoo.com; r-help@r-project.org From: dwinsem...@comcast.net To: rvarad...@jhmi.edu Subject: Re: [R] Double Integration Date: Fri, 2 Jul 2010 20:40:00 -0400 And an adapt() in fCopulae. -- David. On Jul 2, 2010, at 7:06 PM, Ravi Varadhan wrote: There is no package called `integrate', but there is a function called `adaptIntegrate' in the cubature package. Ravi. -Original Message- From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org ] On Behalf Of Christos Argyropoulos Sent: Friday, July 02, 2010 8:41 AM To: sarah_sanche...@yahoo.com; r-help@r-project.org Subject: Re: [R] Double Integration Function adapt in package integrate maybe? Date: Thu, 1 Jul 2010 05:30:25 -0700 From: sarah_sanche...@yahoo.com To: r-help@r-project.org Subject: [R] Double Integration Dear R helpers I am working on the Bi-variate Normal distribution probabilities. I need to double integrate the following function (actually simplified form of bivariate normal distribution) f(x, y) = exp [ - 0.549451 * (x^2 + y^2 - 0.6 * x * y) ] where 2.696 x 3.54 and -1.51 y 1.98 I need to solve something like INTEGRATE (2.696 to 3.54) dx INTEGRATE [(-1.51 to 1.98)] f(x, y) dy I have referred to stats::integrate but it deals with only one variable. This example appears in Internal Credit Risk Model by Michael Ong (page no. 160). Kindly guide. Regards Sarah [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. _ Hotmail: Trusted email with powerful SPAM protection. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. David Winsemius, MD West Hartford, CT _ Hotmail: Free, trusted and rich email service. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] merging plot labels in a lattice plot
Hi: On Fri, Jul 2, 2010 at 8:57 AM, Rajarshi Guha rajarshi.g...@gmail.comwrote: Hi, I have a lattice lot conditioned on two variables. Example code is: library(lattice) x - data.frame(d=runif(100), f1=sample(c('yes', 'no'),100,replace=TRUE), f2=c(rep('Run1',30),rep('Run2',30),rep('Run3',40))) histogram(~d | f1 + f2, x) In the plot, for a given value of f2, there are two panels, one for 'n' and one for 'yes'. But above each panel I get the value of f2. What I'd like to be able to do is to have the value of f2 span the two panels (ie merge the green rows and use a single label). One alternative to Peter's suggestion is to use the strip.combined() function in the Lattice book (p. 197) which merges the two strip labels into one: strip.combined - function(which.given, which.panel, factor.levels, ...) { if (which.given == 1) { panel.rect(0, 0, 1, 1, col = grey90, border = 1) panel.text(x = 0, y = 0.5, pos = 4, lab = factor.levels[which.panel[which.given]]) } if (which.given == 2) { panel.text(x = 1, y = 0.5, pos = 2, lab = factor.levels[which.panel[which.given]]) } } and then call the histogram function as follows: histogram(~ d | f1 + f2, data = x, strip = strip.combined) or histogram(~ d | f1 + f2, data = x, strip = strip.combined, as.table = TRUE) if you prefer the Run* values to go from top to bottom instead. If you'd prefer a different layout for the strip labels but of this general form, you could write your own panel function modeled on the strip.combined() function given above. HTH, Dennis Any pointers as to how I could acheive this would be appreciated Thanks, -- Rajarshi Guha NIH Chemical Genomics Center __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Change the frequency of a ts?
I'm trying to convert a column of a table into a ts object. The data is monthly, so I want the ts frequency to be 12. I did this ... filings.ts = as.ts(Filings.100K, frequency=12) filings.ts Time Series: Start = 1 End = 311 Frequency = 1 [1] 246.9336 305.6789 ... ... tsp(filings.ts) [1] 1 311 1 tsp(filings.ts) - c(1,311,12) Error in attr(x, tsp) - value : invalid time series parameters specified What am I doing wrong here? I can't seem to be able to change the frequency from 1 to 12. Thanks! Nick Frazier [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Change the frequency of a ts?
On Sat, 3 Jul 2010, Nicholas R Frazier wrote: I'm trying to convert a column of a table into a ts object. The data is monthly, so I want the ts frequency to be 12. I did this ... filings.ts = as.ts(Filings.100K, frequency=12) Use the constructor function ts(), not the coercion function as.ts(). The latter does not have a frequency argument. See ?ts. filings.ts Time Series: Start = 1 End = 311 Frequency = 1 [1] 246.9336 305.6789 ... ... tsp(filings.ts) [1] 1 311 1 tsp(filings.ts) - c(1,311,12) Error in attr(x, tsp) - value : invalid time series parameters specified What am I doing wrong here? Not reading the documentation? c(1, 311, 12) are not valid time series properties because it would imply that your series as length 311 * 12 + 1, which is not the case. Z I can't seem to be able to change the frequency from 1 to 12. Thanks! Nick Frazier [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Odp: Problem with aggregating data across time points
Thanks for all your help, that has worked a treat. To answer your questions, I want to include the zero rows because I am going to analyse using mixed models (with dummies for day of week, location etc.) and I thought it was necessary to include a complete list of time variables, but now I'm wondering if it is necessary. As for the empty rows, the database is generated automatically by the incidents reporting system and is a bit of a mess, so I want to make sure that the code doesn't stumble over such things. Thanks again all! On 2 Jul 2010, at 17:14, David Winsemius dwinsem...@comcast.net wrote: On Jul 2, 2010, at 11:55 AM, Petr PIKAL wrote: Hi did you try aggregate? aggregate(data[, 5:8],list(data$Date), sum, na.rm=T) Group.1 verbal self.harm violence_objects violence 1 0 000 2 01/04/07 251539 3 02/04/07 24 68 13 4 03/04/07 17130 10 aggregate(data[, 5:8],list(data$Location,data$Date), sum, na.rm=T) That address his A) request: Here is the application of aggregate to his B) request (I think): # Not e that Date is not of class Date but is rather a factor that includes as a level. aggregate(series[, 5:8],list(series$Date, series$Location), sum, na.rm=T) Group.1 Group.2 verbal self.harm violence_objects violence 10 000 2 A 0 000 3 01/04/07 A 7 103 4 02/04/07 A 8 201 5 03/04/07 A 0 002 6 B 0 000 7 01/04/07 B 3 201 8 02/04/07 B 4 200 9 03/04/07 B 4 003 10C 0 000 11 01/04/07 C 4 232 12 02/04/07 C 0 042 13 03/04/07 C 1 105 14D 0 000 15 01/04/07 D 7 603 16 02/04/07 D 0 009 17 03/04/07 D 41100 18E 0 000 19 01/04/07 E 4 300 20 02/04/07 E 4 040 21 03/04/07 E 8 100 22F 0 000 23 01/04/07 F 0 100 24 02/04/07 F 8 201 So perhaps an output with less extraneous input would be better: with(series[series$Date != , ], aggregate(list(verbal=verbal, self.harm=self.harm, viol_obj=violence_objects, violence=violence), list(Date, Location), sum, na.rm=T) ) Group.1 Group.2 verbal self.harm viol_obj violence 1 01/04/07 A 7 103 2 02/04/07 A 8 201 3 03/04/07 A 0 002 4 01/04/07 B 3 201 5 02/04/07 B 4 200 6 03/04/07 B 4 003 7 01/04/07 C 4 232 8 02/04/07 C 0 042 9 03/04/07 C 1 105 10 01/04/07 D 7 603 11 02/04/07 D 0 009 12 03/04/07 D 41100 13 01/04/07 E 4 300 14 02/04/07 E 4 040 15 03/04/07 E 8 100 16 01/04/07 F 0 100 17 02/04/07 F 8 201 BTW, why do you have empty rows? Regards Petr Hello- I have a dataset which basically looks like this: Location Sex Date Time VerbalSelf harm Violence_objects Violence A 1 1-4-2007 1800 3 0 1 3 A 1 1-4-2007 1230 21 2 4 D 2 2-4-2007 1100 04 0 0 ... I've put a dput of the first section of the data at the end of this email. Basically I have these data for several days across all of the dates, so 2 or more on 1-4-2007, 2 or more on 2-4-2007, and so on
Re: [R] Change the frequency of a ts?
Am 03.07.2010 13:55, schrieb Nicholas R Frazier: I'm trying to convert a column of a table into a ts object. The data is monthly, so I want the ts frequency to be 12. I did this ... filings.ts = as.ts(Filings.100K, frequency=12) try: filings.ts - ts(Filings.100K, frequency=12) example: test-runif(312) test.ts-ts(test, frequency=12) tsp(test.ts) plot(test.ts) Oh I am late, Achim was faster... Cheers Stefan __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Non-exported data sets?
Sure. The code uses objects() to find the exported objects in the package, so I guess the offending object will be there. You can check for yourself by loading the package and calling objects() on the package environment. So I guess my question then is how do data sets and namespaces interact? All data objects are automatically exposed and cannot be controlled through a namespace? Following the hint Two exceptions are allowed: if the R subdirectory contains a file sysdata.rda (a saved image of R objects) this will be lazy-loaded into the name space/package environment – this is intended for system datasets that are not intended to be user-accessible via data. I also tried using sysdata.rda, but the contents still appear to be exported. Hadley -- Assistant Professor / Dobelman Family Junior Chair Department of Statistics / Rice University http://had.co.nz/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] ggplot qplot bar removing bars when truncating scale
This is possible in ggplot2, but it's an not appropriate use of a bar chart - because length is used to convey value, chopping the bottoms of the bars of will give a misleading impression of the data. Instead, use a dot plot: data$Q - unlist(lapply(data$Q, function(x) paste(strwrap(x, 20), collapse = \n))) qplot(mean, Q, data = data, colour = variable, xlab = NULL, ylab = NULL) Hadley On Wed, Jun 30, 2010 at 10:12 AM, ml692787 matthew.lester@gmail.com wrote: I'm having problems with this example, it is posted with reproduceable code below, both with the normal 0-6 scale and the desired 3-6 scale (with bars removed). How can I get the graph to have the desired 3-6 scale without removing the bars. Thanks! #Data mean=as.numeric(c(5.117647059,5,4.947368421,4.85,4.6875,4.545454545,4.473684211,4.470588235,4.428571429,4.08333,3.421052632,3.235294118)) data=as.data.frame(cbind(mean,c(Achievement,Achievement,Achievement,Impact,Achievement,Achievement,Achievement,Impact,Impact,Impact,Impact,Impact),c(Update knowledge and skills,Meet requirements for current position,Discover new job opportunities,Discover new job opportunities,Transition to a new job,Meet requirements for certificaiton,Personal enrichment,Update knowledge and skills,Meet requirements for current position,Meet requirements for certificaiton,Personal enrichment,Transition to a new job))) colnames(data)=c(mean,variable,Q) data[,1]=mean #Plot p=qplot(data=data,data$Q,data$mean,fill=data$variable,geom=bar,stat=identity,position=dodge,binwidth=2,ylab=NULL,xlab=NULL,width=.75) #With 0-6 Scale p + scale_x_discrete(expand=c(0,0)) + scale_y_continuous(limits=c(0,7),breaks=seq(from=0,to=6,by=.5),expand=c(0,0)) + coord_flip() + scale_fill_manual(values=c(darkmagenta,lightgoldenrod1)) + opts( panel.background = theme_rect(colour = NA), panel.background = theme_blank(), panel.grid.minor = theme_blank(), axis.title.x= theme_blank(), axis.title.y= theme_blank(), axis.text.y=theme_text(size=12,hjust=1), legend.text=theme_text(size=14) ) #With 3-6 Scale (Bars Deleted) p + scale_x_discrete(expand=c(0,0)) + scale_y_continuous(limits=c(3,6),breaks=seq(from=3,to=6,by=.5),expand=c(0,0)) + coord_flip() + scale_fill_manual(values=c(darkmagenta,lightgoldenrod1)) + opts( panel.background = theme_rect(colour = NA), panel.background = theme_blank(), panel.grid.minor = theme_blank(), axis.title.x= theme_blank(), axis.title.y= theme_blank(), axis.text.y=theme_text(size=12,hjust=1), legend.text=theme_text(size=14) ) There is probably an option I'm missing or maybe my data should be set up differently, any help would be much appreciated!! -- View this message in context: http://r.789695.n4.nabble.com/ggplot-qplot-bar-removing-bars-when-truncating-scale-tp2272735p2272735.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Assistant Professor / Dobelman Family Junior Chair Department of Statistics / Rice University http://had.co.nz/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Double Integration
Hi Ravi, your suggestion helped me as well a lot. If I look into that function, I see this function is calling another function : .Call(doCubature, as.integer(fDim), body(f.check), as.double(lowerLimit), as.double(upperLimit), as.integer(maxEval), as.double(absError), as.double(tol), new.env(), PACKAGE = cubature) How I can see the interior of this doCubature? Thanks, -Original Message- From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] On Behalf Of Ravi Varadhan Sent: 03 July 2010 04:36 To: 'Christos Argyropoulos'; sarah_sanche...@yahoo.com; r-help@r-project.org Subject: Re: [R] Double Integration There is no package called `integrate', but there is a function called `adaptIntegrate' in the cubature package. Ravi. -Original Message- From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] On Behalf Of Christos Argyropoulos Sent: Friday, July 02, 2010 8:41 AM To: sarah_sanche...@yahoo.com; r-help@r-project.org Subject: Re: [R] Double Integration Function adapt in package integrate maybe? Date: Thu, 1 Jul 2010 05:30:25 -0700 From: sarah_sanche...@yahoo.com To: r-help@r-project.org Subject: [R] Double Integration Dear R helpers I am working on the Bi-variate Normal distribution probabilities. I need to double integrate the following function (actually simplified form of bivariate normal distribution) f(x, y) = exp [ - 0.549451 * (x^2 + y^2 - 0.6 * x * y) ] where 2.696 x 3.54 and -1.51 y 1.98 I need to solve something like INTEGRATE (2.696 to 3.54) dx INTEGRATE [(-1.51 to 1.98)] f(x, y) dy I have referred to stats::integrate but it deals with only one variable. This example appears in Internal Credit Risk Model by Michael Ong (page no. 160). Kindly guide. Regards Sarah [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. _ Hotmail: Trusted email with powerful SPAM protection. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Double Integration
On Sat, 3 Jul 2010, Christos Argyropoulos wrote: There used to be an adapt package with an integrate function (I inverted the function/package name by mistake) in CRAN but it has been removed. Anyone knows why? It lacked a valid licence. It wasn't actually removed, rather archived: see http://cran.r-project.org/src/contrib/Archive/adapt/ Christos CC: argch...@hotmail.com; sarah_sanche...@yahoo.com; r-help@r-project.org From: dwinsem...@comcast.net To: rvarad...@jhmi.edu Subject: Re: [R] Double Integration Date: Fri, 2 Jul 2010 20:40:00 -0400 And an adapt() in fCopulae. -- David. On Jul 2, 2010, at 7:06 PM, Ravi Varadhan wrote: There is no package called `integrate', but there is a function called `adaptIntegrate' in the cubature package. Ravi. -Original Message- From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org ] On Behalf Of Christos Argyropoulos Sent: Friday, July 02, 2010 8:41 AM To: sarah_sanche...@yahoo.com; r-help@r-project.org Subject: Re: [R] Double Integration Function adapt in package integrate maybe? Date: Thu, 1 Jul 2010 05:30:25 -0700 From: sarah_sanche...@yahoo.com To: r-help@r-project.org Subject: [R] Double Integration Dear R helpers I am working on the Bi-variate Normal distribution probabilities. I need to double integrate the following function (actually simplified form of bivariate normal distribution) f(x, y) = exp [ - 0.549451 * (x^2 + y^2 - 0.6 * x * y) ] where 2.696 x 3.54 and -1.51 y 1.98 I need to solve something like INTEGRATE (2.696 to 3.54) dx INTEGRATE [(-1.51 to 1.98)] f(x, y) dy I have referred to stats::integrate but it deals with only one variable. This example appears in Internal Credit Risk Model by Michael Ong (page no. 160). Kindly guide. Regards Sarah [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. _ Hotmail: Trusted email with powerful SPAM protection. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. David Winsemius, MD West Hartford, CT _ Hotmail: Free, trusted and rich email service. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Brian D. Ripley, rip...@stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UKFax: +44 1865 272595 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] logistic regression - glm() - example in Dalgaard's book ISwR
You may find both of Alan Agresti's books on categorcial data analysis useful. Try googling both books and then search the word grouped within each book. Agresti refers to the difference you describe as grouped versus ungrouped data. The likelihoods differ and all summaries based on the likelihood will also differ. On Fri, Jul 2, 2010 at 11:33 PM, Paulo Barata pbar...@infolink.com.br wrote: Dear R-list members, I would like to pose a question about the use and results of the glm() function for logistic regression calculations. The question is based on an example provided on p. 229 in P. Dalgaard, Introductory Statistics with R, 2nd. edition, Springer, 2008. By means of this example, I was trying to practice the different ways of entering data in glm(). In his book, Dalgaard provides data from a case-based study about hypertension summarized in the form of a table. He shows two ways of entering the response (dependent) variable data in glm(): (1) as a matrix of successes/failures (diseased/ healthy); (2) as the proportion of people diseased for each combination of independent variable's categories. I tried to enter the response variable in glm() in a third way: by reconstructing, from the table, the original data in a case-based format, that is, a data frame in which each row shows the data for one person. In this situation, the response variable would be coded as a numeric 0/1 vector, 0=failure, 1=success, and so it would be entered in glm() as a numeric 0/1 vector. The program below presents the calculations for each of the three ways of entering data - the first and second methods were just copied from Dalgaard's book. The three methods produce the same results with regard to the estimated coefficients, when the output is seen with five decimals (although regression 3 seems to have produced slightly different std.errors). My main question is: Why are the residual deviance, its degrees of freedom and the AIC produced by regression 3 completely different when compared to those produced by regressions 1 and 2 (which seem to produce equal results)? It seems that the degrees of freedom in regressions 1 and 2 are based on the size (number of rows) of table d (see the output of the program below), but this table is just a way of summarizing the data. The degrees of freedom in regressions 1 and 2 are not based on the actual number of cases (people) examined, which is n=433. I understand that no matter the way of entering the data in glm(), we are always analyzing the same data, which are those presented in table format on Dalgaard's page 229 (these data are in data.frame d in the program below). So I understand that the three ways of entering data in glm() should produce the same results. Secondarily, why are the std.errors in regression 3 slightly different from those calculated in regressions 1 and 2? I am using R version 2.11.1 on Windows XP. Thank you very much. Paulo Barata ##== begin = ## data in: P. Dalgaard, Introductory Statistics with R, ## 2nd. edition, Springer, 2008 ## logistic regression - example in Dalgaard's Section 13.2, ## page 229 rm(list=ls()) ## data provided on Dalgaard's page 229: no.yes - c(No,Yes) smoking - gl(2,1,8,no.yes) obesity - gl(2,2,8,no.yes) snoring - gl(2,4,8,no.yes) n.tot - c(60,17,8,2,187,85,51,23) n.hyp - c(5,2,1,0,35,13,15,8) d - data.frame(smoking,obesity,snoring,n.tot,n.hyp) ## d is the data to be analyzed, in table format ## d is the first table on Dalgaard's page 229 ## n.tot = total number of cases ## n.hyp = people with hypertension d ## regression 1: Dalgaard's page 230 ## response variable entered in glm() as a matrix of ## successes/failures hyp.tbl - cbind(n.hyp,n.tot-n.hyp) regression1 - glm(hyp.tbl~smoking+obesity+snoring, family=binomial(logit)) ## regression 2: Dalgaard's page 230 ## response variable entered in glm() as proportions prop.hyp - n.hyp/n.tot regression2 - glm(prop.hyp~smoking+obesity+snoring, weights=n.tot,family=binomial(logit)) ## regression 3 (well below): data entered in glm() ## by means of 'reconstructed' variables ## variables with names beginning with 'r' are ## 'reconstructed' from data in data.frame d. ## The objective is to reconstruct the original ## data from which the table on Dalgaard's page 229 ## has been produced rsmoking - c(rep('No',d[1,4]),rep('Yes',d[2,4]), rep('No',d[3,4]),rep('Yes',d[4,4]), rep('No',d[5,4]),rep('Yes',d[6,4]), rep('No',d[7,4]),rep('Yes',d[8,4])) rsmoking - factor(rsmoking) length(rsmoking) # just a check robesity - c(rep('No', d[1,4]),rep('No', d[2,4]), rep('Yes',d[3,4]),rep('Yes',d[4,4]), rep('No', d[5,4]),rep('No', d[6,4]), rep('Yes',d[7,4]),rep('Yes',d[8,4])) robesity - factor(robesity) length(robesity) # just a check
Re: [R] logistic regression - glm() - example in Dalgaard's book ISwR
On Jul 2, 2010, at 11:33 PM, Paulo Barata wrote: Dear R-list members, I would like to pose a question about the use and results of the glm() function for logistic regression calculations. The question is based on an example provided on p. 229 in P. Dalgaard, Introductory Statistics with R, 2nd. edition, Springer, 2008. By means of this example, I was trying to practice the different ways of entering data in glm(). In his book, Dalgaard provides data from a case-based study about hypertension summarized in the form of a table. He shows two ways of entering the response (dependent) variable data in glm(): (1) as a matrix of successes/failures (diseased/ healthy); (2) as the proportion of people diseased for each combination of independent variable's categories. I tried to enter the response variable in glm() in a third way: by reconstructing, from the table, the original data in a case-based format, that is, a data frame in which each row shows the data for one person. In this situation, the response variable would be coded as a numeric 0/1 vector, 0=failure, 1=success, and so it would be entered in glm() as a numeric 0/1 vector. The program below presents the calculations for each of the three ways of entering data - the first and second methods were just copied from Dalgaard's book. The three methods produce the same results with regard to the estimated coefficients, when the output is seen with five decimals (although regression 3 seems to have produced slightly different std.errors). My main question is: Why are the residual deviance, its degrees of freedom and the AIC produced by regression 3 completely different when compared to those produced by regressions 1 and 2 (which seem to produce equal results)? It seems that the degrees of freedom in regressions 1 and 2 are based on the size (number of rows) of table d (see the output of the program below), but this table is just a way of summarizing the data. The degrees of freedom in regressions 1 and 2 are not based on the actual number of cases (people) examined, which is n=433. I first encountered this phenomenon 25 years ago when using GLIM. The answer from my statistical betters was that the deviance is actually only established up to a constant and that it is only differences in deviance that can be properly interpreted. The same situation exists with indefinite integrals in calculus. I understand that no matter the way of entering the data in glm(), we are always analyzing the same data, which are those presented in table format on Dalgaard's page 229 (these data are in data.frame d in the program below). So I understand that the three ways of entering data in glm() should produce the same results. The differences between equivalent nested models should remain the same (up to numerical accuracy). 411.42 on 432 degrees of freedom -398.92 on 429 - 12.5 3 14.1259 on 7 degrees of freedom -1.6184 on 4 -- 12.50753 Secondarily, why are the std.errors in regression 3 slightly different from those calculated in regressions 1 and 2? You mean the differences 4 places to the right of the decimal??? I am using R version 2.11.1 on Windows XP. Thank you very much. Paulo Barata ##== begin = ## data in: P. Dalgaard, Introductory Statistics with R, ## 2nd. edition, Springer, 2008 ## logistic regression - example in Dalgaard's Section 13.2, ## page 229 rm(list=ls()) Personally, I rather annoyed when people post this particular line without commenting it out. It is basically saying that your code is terribly much more important than whatever else might be in a user's workspace. ## data provided on Dalgaard's page 229: no.yes - c(No,Yes) smoking - gl(2,1,8,no.yes) obesity - gl(2,2,8,no.yes) snoring - gl(2,4,8,no.yes) n.tot - c(60,17,8,2,187,85,51,23) n.hyp - c(5,2,1,0,35,13,15,8) d - data.frame(smoking,obesity,snoring,n.tot,n.hyp) ## d is the data to be analyzed, in table format ## d is the first table on Dalgaard's page 229 ## n.tot = total number of cases ## n.hyp = people with hypertension d ## regression 1: Dalgaard's page 230 ## response variable entered in glm() as a matrix of ## successes/failures hyp.tbl - cbind(n.hyp,n.tot-n.hyp) regression1 - glm(hyp.tbl~smoking+obesity+snoring, family=binomial(logit)) ## regression 2: Dalgaard's page 230 ## response variable entered in glm() as proportions prop.hyp - n.hyp/n.tot regression2 - glm(prop.hyp~smoking+obesity+snoring, weights=n.tot,family=binomial(logit)) ## regression 3 (well below): data entered in glm() ## by means of 'reconstructed' variables ## variables with names beginning with 'r' are ## 'reconstructed' from data in data.frame d. ## The objective is to reconstruct the original ## data from which the table on Dalgaard's page 229 ## has been produced rsmoking - c(rep('No',d[1,4]),rep('Yes',d[2,4]),
Re: [R] XML and RCurl: problem with encoding (htmlTreeParse)
Hi Ryusuke I would use the encoding parameter of htmlParse() and download and parse the content in one operation: htmlParse(http://home.sina.com;, encoding = UTF-8) If you want to use getURL() in RCurl, use the .encoding parameter You didn't tell us the output of Sys.getlocale() or how your terminal/console is configured, so the above may vary under your configuration, but works on various machines for me with different settings. D. Ryusuke Kenji wrote: Hi All, First method:- library(XML) theurl - http://home.sina.com; download.file(theurl, tmp.html) txt - readLines(tmp.html) txt - htmlTreeParse(txt, error=function(...){}, useInternalNodes = TRUE) g - xpathSApply(txt, //p, function(x) xmlValue(x)) head(grep( , g, value=T)) [1] ?? | ?? | ENGLISH ??? ??? [3] ??? ?? ??(???) ?? [5] ???? ??! ? ??! ! SecondMethod:- library(RCurl) theurl - getURL(http://home.sina.com,encoding='GB2312') Encoding(theurl) [1]unknown txt - readLines(con=textConnection(theurl),encoding='GB2312') txt[5:10] #show the lines which occurred encoding problem. [1] meta http-equiv=\Content-Type\ content=\text/html; charset=utf-8\ / [2] titleSINA.com US ? -??/title [3] meta name=\Keywords\ content=\, ???, ???, ??,, SINA, US, News, Chinese, Asia\ / [4] meta name=\Description\ content=\???, ???24, , , ??, , ?BBS, ???.\ / [5] [6] link rel=\stylesheet\ type=\text/css\ href=\http://ui.sina.com/assets/css/style_home.css\; / i am trying to read data from a Chinese language website, but the Chinese characters always unreadable, may I know if any good idea to cope such encoding problem in RCurl and XML? Regards, Ryusuke _ [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- There are men who can think no deeper than a fact - Voltaire Duncan Temple Langdun...@wald.ucdavis.edu Department of Statistics work: (530) 752-4782 4210 Mathematical Sciences Bldg. fax: (530) 752-7099 One Shields Ave. University of California at Davis Davis, CA 95616, USA pgpYi9CYtba6H.pgp Description: PGP signature __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] logistic regression - glm() - example in Dalgaard's book ISwR
On Jul 3, 2010, at 9:00 AM, David Winsemius wrote: On Jul 2, 2010, at 11:33 PM, Paulo Barata wrote: Dear R-list members, I would like to pose a question about the use and results of the glm() function for logistic regression calculations. The question is based on an example provided on p. 229 in P. Dalgaard, Introductory Statistics with R, 2nd. edition, Springer, 2008. By means of this example, I was trying to practice the different ways of entering data in glm(). In his book, Dalgaard provides data from a case-based study about hypertension summarized in the form of a table. He shows two ways of entering the response (dependent) variable data in glm(): (1) as a matrix of successes/failures (diseased/ healthy); (2) as the proportion of people diseased for each combination of independent variable's categories. I tried to enter the response variable in glm() in a third way: by reconstructing, from the table, the original data in a case-based format, that is, a data frame in which each row shows the data for one person. In this situation, the response variable would be coded as a numeric 0/1 vector, 0=failure, 1=success, and so it would be entered in glm() as a numeric 0/1 vector. The program below presents the calculations for each of the three ways of entering data - the first and second methods were just copied from Dalgaard's book. The three methods produce the same results with regard to the estimated coefficients, when the output is seen with five decimals (although regression 3 seems to have produced slightly different std.errors). My main question is: Why are the residual deviance, its degrees of freedom and the AIC produced by regression 3 completely different when compared to those produced by regressions 1 and 2 (which seem to produce equal results)? It seems that the degrees of freedom in regressions 1 and 2 are based on the size (number of rows) of table d (see the output of the program below), but this table is just a way of summarizing the data. The degrees of freedom in regressions 1 and 2 are not based on the actual number of cases (people) examined, which is n=433. I first encountered this phenomenon 25 years ago when using GLIM. The answer from my statistical betters was that the deviance is actually only established up to a constant and that it is only differences in deviance that can be properly interpreted. The same situation exists with indefinite integrals in calculus. I understand that no matter the way of entering the data in glm(), we are always analyzing the same data, which are those presented in table format on Dalgaard's page 229 (these data are in data.frame d in the program below). So I understand that the three ways of entering data in glm() should produce the same results. The differences between equivalent nested models should remain the same (up to numerical accuracy). 411.42 on 432 degrees of freedom -398.92 on 429 - 12.5 3 14.1259 on 7 degrees of freedom -1.6184 on 4 -- 12.50753 Secondarily, why are the std.errors in regression 3 slightly different from those calculated in regressions 1 and 2? You mean the differences 4 places to the right of the decimal??? I am using R version 2.11.1 on Windows XP. Thank you very much. Paulo Barata ##== begin = ## data in: P. Dalgaard, Introductory Statistics with R, ## 2nd. edition, Springer, 2008 ## logistic regression - example in Dalgaard's Section 13.2, ## page 229 rm(list=ls()) Personally, I rather annoyed when people post this particular line without commenting it out. It is basically saying that your code is terribly much more important than whatever else might be in a user's workspace. ## data provided on Dalgaard's page 229: no.yes - c(No,Yes) smoking - gl(2,1,8,no.yes) obesity - gl(2,2,8,no.yes) snoring - gl(2,4,8,no.yes) n.tot - c(60,17,8,2,187,85,51,23) n.hyp - c(5,2,1,0,35,13,15,8) d - data.frame(smoking,obesity,snoring,n.tot,n.hyp) ## d is the data to be analyzed, in table format ## d is the first table on Dalgaard's page 229 ## n.tot = total number of cases ## n.hyp = people with hypertension d ## regression 1: Dalgaard's page 230 ## response variable entered in glm() as a matrix of ## successes/failures hyp.tbl - cbind(n.hyp,n.tot-n.hyp) regression1 - glm(hyp.tbl~smoking+obesity+snoring, family=binomial(logit)) ## regression 2: Dalgaard's page 230 ## response variable entered in glm() as proportions prop.hyp - n.hyp/n.tot regression2 - glm(prop.hyp~smoking+obesity+snoring, weights=n.tot,family=binomial(logit)) ## regression 3 (well below): data entered in glm() ## by means of 'reconstructed' variables ## variables with names beginning with 'r' are ## 'reconstructed' from data in data.frame d. ## The objective is to
Re: [R] XML and RCurl: problem with encoding (htmlTreeParse)
Hi Prof, Thank you for your reply. Sorry that I missed out the below information. Sys.getlocale() [1] LC_COLLATE=English_United States.1252;LC_CTYPE=English_United States.1252;LC_MONETARY=English_United States.1252;LC_NUMERIC=C;LC_TIME=English_United States.1252 I have just noticed that traditional chinese character cause the encoding problem, while chinese simplified works fine. library(RCurl) theurl - getURL(http://home.sina.com,encoding='utf8') #Encoding(theurl) #[1]latin1 txt - readLines(con=textConnection(theurl),encoding='utf8') write.table(file='D:/fileas.txt',txt) When I open the fileas.txt, the Chinese traditional character readable in notepad, but when I try to read file to Rgui:- smple - scan('D:/fileas.txt',what='') Then it comes to unrecognisable character again, I was wondering if Rgui support traditional Chinese character now... I think I need to looking for solution of inter-Chinese character's translation. Thank you. Best, Ryusuke === Hi Ryusuke I would use the encoding parameter of htmlParse() and download and parse the content in one operation: htmlParse(http://home.sina.com;, encoding = UTF-8) If you want to use getURL() in RCurl, use the .encoding parameter You didn't tell us the output of Sys.getlocale() or how your terminal/console is configured, so the above may vary under your configuration, but works on various machines for me with different settings. D. Ryusuke Kenji wrote: Hi All, First method:- library(XML) theurl - http://home.sina.com; download.file(theurl, tmp.html) txt - readLines(tmp.html) txt - htmlTreeParse(txt, error=function(...){}, useInternalNodes = TRUE) g - xpathSApply(txt, //p, function(x) xmlValue(x)) head(grep( , g, value=T)) [1] ?? | ?? | ENGLISH ??? ??? [3] ??? ?? ??(???) ?? [5] ???? ??! ? ??! ! SecondMethod:- library(RCurl) theurl - getURL(http://home.sina.com,encoding='GB2312') Encoding(theurl) [1]unknown txt - readLines(con=textConnection(theurl),encoding='GB2312') txt[5:10] #show the lines which occurred encoding problem. [1] meta http-equiv=\Content-Type\ content=\text/html; charset=utf-8\ / [2] titleSINA.com US ? -??/title [3] meta name=\Keywords\ content=\, ???, ???, ??,, SINA, US, News, Chinese, Asia\ / [4] meta name=\Description\ content=\???, ???24, , , ??, , ?BBS, ???.\ / [5] [6] link rel=\stylesheet\ type=\text/css\ href=\http://ui.sina.com/assets/css/style_home.css\; / i am trying to read data from a Chinese language website, but the Chinese characters always unreadable, may I know if any good idea to cope such encoding problem in RCurl and XML? Regards, Ryusuke _ [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- There are men who can think no deeper than a fact - Voltaire Duncan Temple Langdun...@wald.ucdavis.edu Department of Statistics work: (530) 752-4782 4210 Mathematical Sciences Bldg. fax: (530) 752-7099 One Shields Ave. University of California at Davis Davis, CA 95616, USA _ ¥á©`¥ë¤òÒ»À¨¥Á¥§¥Ã¥¯£¡Ëû¤ÎoÁÏ¥á©`¥ë¤â¥×¥í¥Ð¥¤¥À©`¥á©`¥ë¤â¡£ [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Assigning entries to categories
On Sat, 3 Jul 2010, LogLord wrote: Thanks for your help! You are right it is not one-to-one assigned that would be indeed very easy... its more like assigning 1000 entries to 60 categories... Unfortunately, the ?match and ?merge did not help me a lot... I am a newbie to such programming stuff in R. It would be great if you could help me again to set this up. Then you need to observe this: PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. If you provide a _reproducible example_ that properly mimics the features of the problem you need to solve, the chance that someone will either solve it for you or point you in the right direction will be better. [stuff deleted] Charles C. Berry(858) 534-2098 Dept of Family/Preventive Medicine E mailto:cbe...@tajo.ucsd.edu UC San Diego http://famprevmed.ucsd.edu/faculty/cberry/ La Jolla, San Diego 92093-0901 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Double Integration
Bogaso Christofer bogaso.christofer at gmail.com writes: Hi Ravi, your suggestion helped me as well a lot. If I look into that function, I see this function is calling another function : .Call(doCubature, as.integer(fDim), body(f.check), as.double(lowerLimit), as.double(upperLimit), as.integer(maxEval), as.double(absError), as.double(tol), new.env(), PACKAGE = cubature) How I can see the interior of this doCubature? Find the original code for the 'cubature' package at http://ab-initio.mit.edu/wiki/index.php/Cubature plus information why the 'adapt' package had to be abandoned and that 'cubature' is based on the same original algorithm of Genz and Malik, but using free and GPLed software. We should not bemoan the loss of the 'adapt' package, 'cubature' and 'R2cuba' are worthy successors for adaptive quadrature. Hans Werner __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Best way to compute a sum
On Fri, 2010-07-02 at 21:23 -0700, Roger Deangelis wrote: Although it does not apply to your series and is impractical, it seems to me that the most accurate algorithm might be to add all the rational numbers whose sum and components can be represented without error in binary first, ie 2.5 + .5 or 1/16 + 1/16 + 1/8. You could also get very clever and investigate a sum that should have an exact binary representation when the individual components do not, ie .1 + .2 + .2 = .5 and correct the sum. Roger Roger I think you must read: What Every Computer Scientist Should Know About Floating-Point Arithmetic ( http://docs.sun.com/source/806-3568/ncg_goldberg.html ) I think your question and others like this question is answer in this paper -- Bernardo Rangel Tura, M.D,MPH,Ph.D National Institute of Cardiology Brazil __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Change the frequency of a ts?
Thanks, Stefan. I'm sure the difference between ts() and as.ts() seemed simple to everyone on the list, but I'd been staring at the help files for a long time and never made the connection. ts's make more sense now. Nick Frazier On Sat, Jul 3, 2010 at 8:10 AM, Stefan Grosse singularit...@gmx.net wrote: Am 03.07.2010 13:55, schrieb Nicholas R Frazier: I'm trying to convert a column of a table into a ts object. The data is monthly, so I want the ts frequency to be 12. I did this ... filings.ts = as.ts(Filings.100K, frequency=12) try: filings.ts - ts(Filings.100K, frequency=12) example: test-runif(312) test.ts-ts(test, frequency=12) tsp(test.ts) plot(test.ts) Oh I am late, Achim was faster... Cheers Stefan [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] How to generate longitudinal data using R
How to generate the longitudinal data with correlation structure of independent , exchangeable and AR (1) through errors? Can someone provide some sample codes? Great appreciation! Thanks much, Yi [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] PDFfontNumber bugs in devPS.c (Re: plain text in Chinese can not be set)
Hi Thanks very much for the report, diagnosis, and patch! I have implemented your fix in the development version of R. Paul Jinsong Zhao wrote: On 2010-7-1 15:24, Jinsong Zhao wrote: Read the source again more carefully. I think I get the solution: Change the following line in PDFfontNumber function in devPS.c: num = 1000 + (cidfontIndex - 1)*5 + 1 + face; to num = 1000 + (cidfontIndex - 1)*5 + face; It appears two times in the function. However, I don't know how to compile the whole R distribution on Windows platform. Would anyone here like to give a test? Thanks in advance! Regards, Jinsong I have compiled R 2.11.1 on a Linux machine, and confirmed that PDFfontNumber function in devPS.c (grDevices library) has a bug, which causes the plain face of CID fonts cannot be accessed when CID fonts were used together with default font family in pdf(). the following is the patch. --- devPS_orig.cSun Apr 25 06:10:04 2010 +++ devPS.c Fri Jul 02 09:46:55 2010 @@ -7267,7 +7267,7 @@ * Use very high font number for CID fonts to avoid * Type 1 fonts */ - num = 1000 + (cidfontIndex - 1)*5 + 1 + face; + num = 1000 + (cidfontIndex - 1)*5 + face; else { /* * Check whether the font is loaded and, if not, @@ -7303,7 +7303,7 @@ } else /* (isCIDFont(family, PDFFonts)) */ { if (addPDFDeviceCIDfont(cidfontfamily, pd, cidfontIndex)) { - num = 1000 + (cidfontIndex - 1)*5 + 1 + face; + num = 1000 + (cidfontIndex - 1)*5 + face; } else { cidfontfamily = NULL; Regards, Jinsong -- Dr Paul Murrell Department of Statistics The University of Auckland Private Bag 92019 Auckland New Zealand 64 9 3737599 x85392 p...@stat.auckland.ac.nz http://www.stat.auckland.ac.nz/~paul/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] interval censored grouped data
Hi All, I have data in the following format: Inspection #failures Month/Year 01/99 5 02/99 20 06/993 01/02 3 for 11 years ... the prob of failure on demand per month pfd is #Total failures / sample size(=total components * 11yrs * 12months). I am required to cross-check pfd by other means. Can I fit a weibull or lognormal plot to this count data and find the failure probability or reliability from this data? exact failure times are not known..only the number of failures in the month is known. please suggest an appropriate method / R code. Thanks, Tims [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] How to generate longitudinal data using R
On Sat, 3 Jul 2010, ZZY ZYBOYS wrote: How to generate the longitudinal data with correlation structure of independent , exchangeable and AR (1) through errors? Can someone provide some sample codes? See http://cran.r-project.org/web/views/Distributions.html Great appreciation! Thanks much, Yi [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Charles C. Berry(858) 534-2098 Dept of Family/Preventive Medicine E mailto:cbe...@tajo.ucsd.edu UC San Diego http://famprevmed.ucsd.edu/faculty/cberry/ La Jolla, San Diego 92093-0901 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] help with predict.lda
HI, Dear community, I am using the linear discriminant analysis to build model and make new predictions: dim(train) #training data [1] 1272 22 dim(valid) # validation data [1] 140 22 lda.fit - lda(out ~ ., data=train, na.action=na.omit, CV=TRUE) # model fitting of linear discriminant analysis on training data predict(lda.fit, valid) # make prediction on validation data Error in UseMethod(predict) : no applicable method for 'predict' applied to an object of class list Can anyone help with this? Thanks so much! -- Sincerely, Changbin -- [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.